
#Importe de librerías
#Manejo de datos
import pandas as pd
import seaborn as sns
import numpy as np
import collections, os
from collections import Counter
from pandas_profiling import ProfileReport
from functools import reduce

#Progress bar
from tqdm import tqdm

import re, unicodedata, spacy
#, es_core_news_sm

#Directory listing
from os import listdir
from os.path import isfile, join

#Texto 
from wordcloud import WordCloud, STOPWORDS
from nltk.corpus import stopwords
stopwords = set(stopwords.words('english'))
from sklearn.preprocessing import MinMaxScaler
from difflib import SequenceMatcher

#Fecha
import dateparser
from datetime import datetime
sns.set()
date_fmt = '%b %Y'

#Graficación
from plotly.subplots import make_subplots
import plotly.graph_objs as go
import matplotlib.pyplot as plt


def txt_preproc(input_str):
    '''
    Función para preprocesar cadenas de texto
    
    Tomado/basado en:
    https://stackoverflow.com/questions/517923/what-is-the-best-way-to-remove-accents-normalize-in-a-python-unicode-string
    https://stackoverflow.com/questions/5843518/remove-all-special-characters-punctuation-and-spaces-from-string
    
    Recibe una cadena de texto en formato string llamado ``input_str``
    
    Primero remueve caracteres especiales
    Remueve acentos y caracteres unicode
    Finalmente con el método .to_lower() convierte los caracteres a minúscula
    
    Args:
        input_str (string):   Cadena de texto a preprocesar
    Returns:
        nfkd_form (string):  Cadena de texto preprocesada.
    '''
    
    nfkd_form = re.sub(r'[?|$|.|!]',r'',input_str) #Remove special chars
    nfkd_form = unicodedata.normalize('NFKD', nfkd_form) #Remove accents
    nfkd_form = u"".join([c for c in nfkd_form if not unicodedata.combining(c)]) #Remove unicode chars
    nfkd_form = (re.sub(r'[^a-zA-Z0-9 ]',r'',nfkd_form)).lower() #Remove missing special chars and convert to lowercase
    
    return nfkd_form


mypath = './dataset/soa_sports_data.csv'

soa_df = pd.read_csv(mypath, encoding='latin1', error_bad_lines=False, low_memory=False)
soa_df.columns = [x.lower() for x in soa_df.columns]

soa_df.head()


soa_df.insert(7, 'titulo_proc', None)
soa_df.insert(8, 'abstract_proc', None)
soa_df.insert(9, 'roe_proc', None)
soa_df.insert(10, 'len_titulo', None)
soa_df.insert(11, 'len_abstract', None)

#Preprocessing text
for row in tqdm(range(len(soa_df))):
    df_title = soa_df['titulo'].iloc[row]
    df_abstract = soa_df['abstract'].iloc[row]
    df_roe = soa_df['revista_or_evento'].iloc[row]
    len_title = len(df_title.split(" "))
    len_abstract = len(df_abstract.split(" "))

    df_title = txt_preproc(df_title)
    df_abstract = txt_preproc(df_abstract)
    df_roe = txt_preproc(df_roe)
    #============================================
    soa_df['titulo_proc'].iloc[row] = df_title
    soa_df['abstract_proc'].iloc[row] = df_abstract
    soa_df['roe_proc'].iloc[row] = df_roe
    soa_df['len_titulo'].iloc[row] = len_title
    soa_df['len_abstract'].iloc[row] = len_abstract

  0%|                                                                                                                    | 0/102 [00:00<?, ?it/s]C:\Users\user\anaconda3\envs\py38dane\lib\site-packages\pandas\core\indexing.py:1637: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  self._setitem_single_block(indexer, value, name)
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████| 102/102 [00:00<00:00, 552.74it/s]


soa_df.head()


soa_df.to_csv('./dataset/soa_sports_data_proc.csv', index=False)


all_year_base_gb = soa_df.groupby(["año"]).count()
all_year_base_gb = all_year_base_gb.reset_index()
all_year_base_gb = all_year_base_gb[["año", "query"]]
all_year_base_gb = all_year_base_gb.rename(columns={'año': 'year', 'query': 'count'})
all_year_base_gb = all_year_base_gb.sort_values(by=['year'])

all_year_base_gb


year_labels = [2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021]

year_values = [int(all_year_base_gb[all_year_base_gb['year']==year_labels[0]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[1]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[2]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[3]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[4]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[5]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[6]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[7]]['count']), 
               int(all_year_base_gb[all_year_base_gb['year']==year_labels[8]]['count'])]

fig = go.Figure(data=[go.Pie(labels=year_labels, values=year_values, hole=.3)])

fig.update_layout(
    title_text="2013 - 2021: Annual Production, Total %",
    legend_title="Year",
    # Add annotations in the center of the donut pies.
    annotations=[dict(text='Prod %', x=0.5, y=0.5, font_size=13, showarrow=False)],
    autosize=False,
    width=950,
    height=400,
    margin=dict(l=0, r=0, t=50, b=0),
    font=dict(size=12))
fig.show()


all_year_base_pt_gb = soa_df.groupby(['año', 'type']).count()

all_year_base_pt_gb = all_year_base_pt_gb.reset_index()

all_year_base_pt_gb = all_year_base_pt_gb[["año", "type", "query"]]
all_year_base_pt_gb = all_year_base_pt_gb.rename(columns={'año': 'year', 'query': 'count'})
all_year_base_pt_gb = all_year_base_pt_gb.sort_values(by=['year'])

all_year_base_pt_gb_arts = all_year_base_pt_gb[all_year_base_pt_gb['type']=='Article']
all_year_base_pt_gb_thes = all_year_base_pt_gb[all_year_base_pt_gb['type']=='Thesis']
all_year_base_pt_gb_book = all_year_base_pt_gb[all_year_base_pt_gb['type']=='Book']
                                                          
all_year_base_pt_gb


#animals=['giraffes', 'orangutans', 'monkeys']
year_labels = ['2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020', '2021']

fig = go.Figure(data=[go.Bar(name='Articles', x=year_labels, y=list(all_year_base_pt_gb_arts['count']), width=0.3),
                      go.Bar(name='Thesis', x=year_labels, y=list(all_year_base_pt_gb_thes['count']), width=0.3), 
                      go.Bar(name='Books', x=year_labels, y=list(all_year_base_pt_gb_book['count']), width=0.3)
])
# Change the bar mode
fig.update_layout(barmode='stack')
fig.update_layout(
    title_text="2013 - 2021: Annual Production by Type",
    xaxis_title="Year",
    yaxis_title="Total",
    legend_title="Type",
    autosize=False,
    width=950,
    height=400,
    margin=dict(l=0, r=0, t=50, b=0),
    font=dict(size=11))
fig.show()


all_year_base_cites_gb = soa_df.groupby(['año', 'citas']).count()
all_year_base_cites_gb = all_year_base_cites_gb.reset_index()

all_year_base_cites_gb = all_year_base_cites_gb[["año", "citas"]]
all_year_base_cites_gb = all_year_base_cites_gb.rename(columns={'año': 'year', 'citas': 'cites'})

all_year_base_cites_gb.head()


df_cites_sum = all_year_base_cites_gb.groupby(['year']).sum().reset_index().rename(columns={'cites': 'tot'})
df_cites_mean = all_year_base_cites_gb.groupby(['year']).mean().reset_index().rename(columns={'cites': 'mean_val'})
df_cites_median = all_year_base_cites_gb.groupby(['year']).median().reset_index().rename(columns={'cites': 'median_val'})
df_cites_max = all_year_base_cites_gb.groupby(['year']).max().reset_index().rename(columns={'cites': 'max_val'})
df_cites_min = all_year_base_cites_gb.groupby(['year']).min().reset_index().rename(columns={'cites': 'min_val'})

cites_data_frames = [df_cites_sum, df_cites_mean, df_cites_median, df_cites_max, df_cites_min]

df_merged = reduce(lambda  left,right: pd.merge(left,right,on=['year'],
                                            how='outer'), cites_data_frames)

df_merged


fig = go.Figure()

year_labels = ['2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020', '2021']

fig.add_trace(go.Bar(x=year_labels,y=list(df_merged.tot), width=0.3, name='Total'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_merged.mean_val), name='Mean'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_merged.median_val), name='Median'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_merged.max_val), name='Max'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_merged.min_val), name='Min'))

fig.update_layout(
    title_text="2013 - 2021: Citations stats",
    xaxis_title="Year",
    yaxis_title="Total",
    legend_title="Stats",
    autosize=False,
    width=950,
    height=400,
    margin=dict(l=0, r=0, t=50, b=0),
    font=dict(size=11))
fig.show()


#Año, citas, len de title, len de abstract
all_year_base_tot_gb = soa_df.groupby(['año', 'citas'], as_index=False).sum()
all_year_base_tot_gb = all_year_base_tot_gb[['año', 'citas', 'len_titulo', 'len_abstract']]
all_year_base_tot_gb = all_year_base_tot_gb.rename(columns={'año': 'year', 'citas': 'cites'})
all_year_base_tot_gb.head()


fig = go.Figure(data=[go.Scatter3d(
    x=all_year_base_tot_gb.len_abstract,
    y=all_year_base_tot_gb.len_titulo,
    z=all_year_base_tot_gb.cites,
    name= 'Color scale',
    mode='markers',
    marker=dict(
        size=8,
        color=all_year_base_tot_gb.year,                # set color to an array/list of desired values
        #colorscale='Viridis',   # choose a colorscale
        opacity=0.8,
        showscale=True,
        colorbar = dict(len=0.8)
    )
)])

# tight layout
#fig.update_layout(margin=dict(l=0, r=0, b=0, t=0))
fig.update_layout(scene = dict(
                    xaxis_title='Abstract lenght',
                    yaxis_title='Title lenght',
                    zaxis_title='Cites'),
                    width=700,
                    margin=dict(r=0, b=0, l=0, t=0))

fig.update_layout(showlegend=True)
#fig.update_traces(legendgroup=all_year_base_tot_gb.year)
fig.update_traces(legendgroup=8)
fig.show()


title_words = ''
title_list_full, title_list_filtered = [], []

# iterate through the csv file 
for val in soa_df.titulo_proc:
    # typecaste each val to string 
    val = str(val) 
    # split the value 
    tokens = val.split()
    
    for tok in tokens:
        title_list_full.append(tok)
        if tok not in stopwords:
            title_list_filtered.append(tok)
        
    title_words += " ".join(tokens)+" "

wordcloud = WordCloud(width = 1000, height = 400, 
                background_color ='white', 
                stopwords = stopwords, 
                min_font_size = 10).generate(title_words) 
  
# plot the WordCloud image                        
plt.figure(figsize = (9, 9), facecolor = None) 
plt.imshow(wordcloud)
plt.axis("off")
plt.tight_layout(pad = 0)

plt.show()


sorted_tc_dict = sorted(dict(Counter(title_list_filtered)).items(), key=lambda kv: kv[1])
title_array = np.asarray(sorted_tc_dict)

labels = np.flipud(title_array[:,0])
values = np.flipud(title_array[:,1]).astype(int)

data = [go.Bar(
   x = labels[:100],
   y = values[:100]
)]
fig = go.Figure(data=data)
fig.update_layout(xaxis_tickangle=-45)
fig.update_layout(
    title_text='Top 100 high-frequency words for titles',
    xaxis_title="Words",
    yaxis_title="Frequency",
)
fig.show()


### Wordcloud review by abstract


abstract_words = ''
abstract_list_full, abstract_list_filtered = [], []

# iterate through the csv file 
for val in soa_df.abstract_proc:
    # typecaste each val to string 
    val = str(val) 
    # split the value 
    tokens = val.split()
    
    for tok in tokens:
        abstract_list_full.append(tok)
        if tok not in stopwords:
            abstract_list_filtered.append(tok)
        
    abstract_words += " ".join(tokens)+" "

wordcloud = WordCloud(width = 1000, height = 400, 
                background_color ='white', 
                stopwords = stopwords, 
                min_font_size = 10).generate(abstract_words) 
  
# plot the WordCloud image                        
plt.figure(figsize = (9, 9), facecolor = None) 
plt.imshow(wordcloud)
plt.axis("off")
plt.tight_layout(pad = 0)

plt.show()


sorted_tc_dict = sorted(dict(Counter(abstract_list_filtered)).items(), key=lambda kv: kv[1])
abstract_array = np.asarray(sorted_tc_dict)

labels = np.flipud(abstract_array[:,0])
values = np.flipud(abstract_array[:,1]).astype(int)

data = [go.Bar(
   x = labels[:100],
   y = values[:100]
)]
fig = go.Figure(data=data)
fig.update_layout(xaxis_tickangle=-45)
fig.update_layout(
    title_text='Top 100 high-frequency words for abstracts',
    xaxis_title="Words",
    yaxis_title="Frequency",
)
fig.show()


df_autors = soa_df.loc[:,'a_1':'a_14']
df_yc = soa_df[['año','citas', 'a_1']]

df_authors_full = reduce(lambda  left,right: pd.merge(left,right,on=['a_1'],
                                            how='outer'), [df_yc, df_autors])

df_authors_full = df_authors_full.rename(columns={'año': 'year', 'citas': 'cites'})
df_authors_full['tot_authors'] = None

all_authors_list = []

for row in tqdm(range(len(df_authors_full))):
    author_row = df_authors_full.loc[row, 'a_1':'a_14'].values.tolist()
    #============================================
    df_authors_full['tot_authors'].iloc[row] = 14-author_row.count('-')
    
    for el in author_row:
        if el!='-':
            all_authors_list.append(el)
        else:
            pass

df_authors_full

  0%|                                                                                                                    | 0/130 [00:00<?, ?it/s]C:\Users\user\anaconda3\envs\py38dane\lib\site-packages\pandas\core\indexing.py:1637: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

100%|████████████████████████████████████████████████████████████████████████████████████████████████████████| 130/130 [00:00<00:00, 1481.23it/s]


df_authors_full_yt = df_authors_full[['year', 'tot_authors']]
df_authors_full_yt = df_authors_full_yt.groupby(['year', 'tot_authors']).count()
df_authors_full_yt = df_authors_full_yt.reset_index()
df_authors_full_yt


df_auth_sum = df_authors_full_yt.groupby(['year']).sum().reset_index().rename(columns={'tot_authors': 'tot'})
df_auth_mean = df_authors_full_yt.groupby(['year']).mean().reset_index().rename(columns={'tot_authors': 'mean_val'})
df_auth_median = df_authors_full_yt.groupby(['year']).median().reset_index().rename(columns={'tot_authors': 'median_val'})
df_auth_max = df_authors_full_yt.groupby(['year']).max().reset_index().rename(columns={'tot_authors': 'max_val'})
df_auth_min = df_authors_full_yt.groupby(['year']).min().reset_index().rename(columns={'tot_authors': 'min_val'})

auth_data_frames = [df_auth_sum, df_auth_mean, df_auth_median, df_auth_max, df_auth_min]

df_auth_merged = reduce(lambda  left,right: pd.merge(left,right,on=['year'],
                                            how='outer'), auth_data_frames)

df_auth_merged


fig = go.Figure()

year_labels = ['2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020', '2021']

fig.add_trace(go.Bar(x=year_labels,y=list(df_auth_merged.tot), width=0.3, name='Total'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_auth_merged.mean_val), name='Mean'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_auth_merged.median_val), name='Median'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_auth_merged.max_val), name='Max'))
fig.add_trace(go.Scatter(x=year_labels,y=list(df_auth_merged.min_val), name='Min'))

fig.update_layout(
    title_text="2013 - 2021: Authors stats",
    xaxis_title="Year",
    yaxis_title="Total",
    legend_title="Stats",
    autosize=False,
    width=950,
    height=400,
    margin=dict(l=0, r=0, t=50, b=0),
    font=dict(size=11))
fig.show()


letter_counts = Counter(all_authors_list)
df_authors = pd.DataFrame.from_dict(letter_counts, orient='index').rename(columns={'index': 'author', 0: 'count'})
df_authors = df_authors.sort_values(by=['count'], ascending=True).reset_index().rename(columns={'index': 'author', 0: 'count'})

fig = go.Figure(go.Bar(
            x=list(df_authors['count']),
            y=list(df_authors['author']),
            orientation='h'))

fig.update_layout(
    title_text="2013 - 2021: Authors stats - Tot. prod.",
    xaxis_title="Total",
    yaxis_title="Author name")

fig.show()

	query	año	citas	type	titulo	abstract	revista_or_evento	a_1	a_2	a_3	...	a_5	a_6	a_7	a_8	a_9	a_10	a_11	a_12	a_13	a_14
0	"soccer" and "analytics" and "world"	2014	89	Article	Feature-driven visual analytics of soccer data	Soccer is one the most popular sports today an...	2014 IEEE conference on visual analytics scien...	Sacha, D	Stein, M.	Schreck, T	...	Deussen, O.	-	-	-	-	-	-	-	-	-
1	"soccer" and "analytics" and "world"	2015	35	Article	Visual Soccer Analytics: Understanding the Cha...	With recent advances in sensor technologies, l...	ISPRS International Journal of Geo-Information	Stein, M.	Häußler, J.	Jäckle, D.	...	Schreck, T.	Keim, D.	-	-	-	-	-	-	-	-
2	"soccer" and "analytics" and "world"	2013	35	Thesis	Machine Learning for Soccer Analytics	Sports analytics has been successfully applied...	University of Leuven	Kumar, G.	-	-	...	-	-	-	-	-	-	-	-	-	-
3	"soccer" and "analytics" and "world"	2014	64	Article	Bagadus: An integrated real-time system for so...	The importance of winning has increased the ro...	ACM Transactions on Multimedia Computing, Comm...	Stensland, H. K.	Gaddam, V. R	Tennøe, M.	...	Næss, M.	Alstad, H. K.	Mortensen, A.	Langseth, R.	Ljødal, S.	Landsverk, Ø.	Griwodz, C.	Halvorsen, P.	Stenhaug, M.	Johansen, D.
4	"soccer" and "analytics" and "world"	2014	0	Article	World Cup sets US live streaming record	The world cup has become the most streamed liv...	FGV Digital Respository	Bond, S.	-	-	...	-	-	-	-	-	-	-	-	-	-

	query	año	citas	type	titulo	abstract	revista_or_evento	titulo_proc	abstract_proc	roe_proc	...	a_5	a_6	a_7	a_8	a_9	a_10	a_11	a_12	a_13	a_14
0	"soccer" and "analytics" and "world"	2014	89	Article	Feature-driven visual analytics of soccer data	Soccer is one the most popular sports today an...	2014 IEEE conference on visual analytics scien...	featuredriven visual analytics of soccer data	soccer is one the most popular sports today an...	2014 ieee conference on visual analytics scien...	...	Deussen, O.	-	-	-	-	-	-	-	-	-
1	"soccer" and "analytics" and "world"	2015	35	Article	Visual Soccer Analytics: Understanding the Cha...	With recent advances in sensor technologies, l...	ISPRS International Journal of Geo-Information	visual soccer analytics understanding the char...	with recent advances in sensor technologies la...	isprs international journal of geoinformation	...	Schreck, T.	Keim, D.	-	-	-	-	-	-	-	-
2	"soccer" and "analytics" and "world"	2013	35	Thesis	Machine Learning for Soccer Analytics	Sports analytics has been successfully applied...	University of Leuven	machine learning for soccer analytics	sports analytics has been successfully applied...	university of leuven	...	-	-	-	-	-	-	-	-	-	-
3	"soccer" and "analytics" and "world"	2014	64	Article	Bagadus: An integrated real-time system for so...	The importance of winning has increased the ro...	ACM Transactions on Multimedia Computing, Comm...	bagadus an integrated realtime system for socc...	the importance of winning has increased the ro...	acm transactions on multimedia computing commu...	...	Næss, M.	Alstad, H. K.	Mortensen, A.	Langseth, R.	Ljødal, S.	Landsverk, Ø.	Griwodz, C.	Halvorsen, P.	Stenhaug, M.	Johansen, D.
4	"soccer" and "analytics" and "world"	2014	0	Article	World Cup sets US live streaming record	The world cup has become the most streamed liv...	FGV Digital Respository	world cup sets us live streaming record	the world cup has become the most streamed liv...	fgv digital respository	...	-	-	-	-	-	-	-	-	-	-

	year	tot	mean_val	median_val	max_val	min_val
0	2013	1874	208.222222	35.0	1177	6
1	2014	750	41.666667	26.5	177	0
2	2015	532	26.600000	22.5	77	0
3	2016	746	93.250000	39.5	306	2
4	2017	156	19.500000	16.5	54	5
5	2018	202	22.444444	24.0	43	1
6	2019	109	21.800000	12.0	80	1
7	2020	22	7.333333	3.0	19	0
8	2021	1	1.000000	1.0	1	1

	year	cites	a_1	a_2	a_3	a_4	a_5	a_6	a_7	a_8	a_9	a_10	a_11	a_12	a_13	a_14	tot_authors
0	2014	89	Sacha, D	Stein, M.	Schreck, T	Keim, D.	Deussen, O.	-	-	-	-	-	-	-	-	-	5
1	2015	35	Stein, M.	Häußler, J.	Jäckle, D.	Janetzko, H.	Schreck, T.	Keim, D.	-	-	-	-	-	-	-	-	6
2	2013	35	Kumar, G.	-	-	-	-	-	-	-	-	-	-	-	-	-	1
3	2014	64	Stensland, H. K.	Gaddam, V. R	Tennøe, M.	Helgedagsrud, E.	Næss, M.	Alstad, H. K.	Mortensen, A.	Langseth, R.	Ljødal, S.	Landsverk, Ø.	Griwodz, C.	Halvorsen, P.	Stenhaug, M.	Johansen, D.	14
4	2014	0	Bond, S.	-	-	-	-	-	-	-	-	-	-	-	-	-	1
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
125	2015	2	Brymer, R.	Holcomb, T. R.	Rodenberg, R. M.	-	-	-	-	-	-	-	-	-	-	-	3
126	2013	33	Drachen, A.	Schubert, M.	-	-	-	-	-	-	-	-	-	-	-	-	2
127	2019	3	Yi?it, A. T.	Samak, B.	Kaya, T.	-	-	-	-	-	-	-	-	-	-	-	3
128	2019	13	Singh, P.	Lamba, P. S.	-	-	-	-	-	-	-	-	-	-	-	-	2
129	2019	80	Baboota, R.	Kaur, H.	-	-	-	-	-	-	-	-	-	-	-	-	2

1. Libraries¶

Functions¶

2. Data loading¶

2.1. Data processing¶

3. Some stats¶

Yearly¶

3.1. Production %¶

3.2. Production by type¶

3.3. Citations¶

3.4. Citations acording title and abstract lenght¶

Wordcloud review by title¶

	year	tot	mean_val	median_val	max_val	min_val
0	2013	15	3.000000	3.0	5	1
1	2014	46	5.750000	4.5	14	1
2	2015	28	4.000000	4.0	7	1
3	2016	15	3.750000	3.5	6	2
4	2017	24	4.000000	3.5	8	1
5	2018	14	3.500000	3.5	5	2
6	2019	10	3.333333	3.0	5	2
7	2020	20	6.666667	7.0	9	4
8	2021	3	3.000000	3.0	3	3